Filtered Manifold Alignment

TL;DR

This paper introduces Filtered Manifold Alignment (FMA), a semi-supervised technique for domain adaptation that reduces computational complexity and improves classification accuracy across diverse image datasets.

Contribution

The paper proposes a novel FMA method that efficiently aligns disparate domains in low-dimensional spaces, advancing domain adaptation techniques.

Findings

FMA achieves state-of-the-art accuracy on benchmark image datasets.

FMA reduces computational complexity compared to previous methods.

FMA effectively aligns domains with different feature sets.

Abstract

Domain adaptation is an essential task in transfer learning to leverage data in one domain to bolster learning in another domain. In this paper, we present a new semi-supervised manifold alignment technique based on a two-step approach of projecting and filtering the source and target domains to low dimensional spaces followed by joining the two spaces. Our proposed approach, filtered manifold alignment (FMA), reduces the computational complexity of previous manifold alignment techniques, is flexible enough to align domains with completely disparate sets of feature and demonstrates state-of-the-art classification accuracy on multiple benchmark domain adaptation tasks composed of classifying real world image datasets.